新的普适计算环境下上下文缓存置换算法

由于普适计算环境高度动态的特点以及无线环境连接易中断、传输速度低等方面的约束，使得普适计算应用对于上下文的访问开销非常大。为此，首先给出一个上下文感知系统框架，然后提出了一种基于规则的上下文缓存置换算法--RCRA，算法根据上下文的被访问概率、上下文时效性及历史访问次数决定是否将其置换出缓存。当有新的上下文需要进入缓存时执行该算法，以保证缓存中的上下文最新且最有价值。实验表明，RCRA不仅在命中率方面有较大提高，而且能够有效降低上下文访问的开销。算法应用于基于推理规则的上下文感知系统中，具有良好的可实用性。

New context caching replacement algorithm in ubiquitous computing

Due to the high dynamics of pervasive computing environment and the constraints of easy interruption and low transmission rate of wireless network, the overhead of context access is very huge. To solve these problems, a framework of context-aware system was proposed in this paper, and then a context caching replacement algorithm based on rules (RCRA) was introduced. The proposed algorithm determined whether to replace a context in the cache based on its access probability, timeliness and access history. When a new context was to get into the context cache, the algorithm was used to ensure that the latest and the most valuable context stayed in the cache. The experimental results show that the RCRA improves the hit rate and effectively reduces the overhead of context access. The RCRA is used in the rules-based context-aware system, and the algorithm has good utility.